It’s taken over 20 years for e-commerce to get to 16% of global retail sales. If we draw a parallel, agentic commerce will be much faster. I predict we’ll get to 10% in three to five years.
The mass consumer adoption of AI technologies, whether that’s OpenAI’s ChatGPT with 800 million weekly active users or Google AI with 1.5 billion monthly active users, means there’s a huge, scaled audience already for agentic commerce.
Brands are recognising that these are new interfaces within which consumers who are researching for goods and services might eventually want to fulfil the whole, end-to-end buying journey.
As with the beginning of e-commerce, AI represents a new distribution channel for businesses to get their heads around, but it’s easier because the audience is already built, there are already five and a half billion people online, and the guardrails of commerce.
If you ask anyone, “In five to 10 years, do you think there will be more buying through AI than today?” Very few people are saying ‘no’. We all know it’s coming; it’s just a question of how fast and how scaled it is. Is it a micro-niche of really confident people, or is it a mass market proposition?
The challenge with LLMs is that, very quickly, as in many digital markets, we see a ‘winner takes most’ kind of market. Just think: who’s number two to Amazon? It would be hard to name someone. If you give up that middle ground and lose the ability to become the answer to a particular style of question that a consumer might have, it might be very hard to wrestle back that ground later on.
At the moment, it’s all about exploring the existing AI territories and putting something there. Many businesses want to get out there, not just because it shows to the street that they’re AI-forward and they’re thinking about these things, but, in the medium term, because there’s also true revenue to be won.
If you buy into the narrative that there’s going to be more AI in the world rather than less in the future, it’s very important to get your learnings in early. If you start developing an agentic commerce option on an LLM platform, the lessons you learn allow you to build the second and third generations. If you decide to sit it out for two or three years, then, when you enter, you’re entering at generation zero, and you still have to go through all those steps to get those learnings and improve what you have.
Part of the thing about going early is that it’s less risky. This is because one of the key byproducts of AI is the data that AI generates, which is data you wouldn’t otherwise have, and it can inform your development cycle to improve and get faster. When people think about AI, they forget that the ‘exhaust’ is the real gold, the exhaust being all this data coming out. It can help with answering the questions you’re really bad at answering: what are the things that generate a sale and don’t generate a sale? Unless you’ve got some way to capture and understand that information, you can’t see what’s happening. All you can see is the headline growth, and you do not understand the clues along the way
As a brand, if you decide to wait until agentic commerce has scaled before joining, you’re starting from square one and don’t already have an optimised process of how you manage these things, how to handle returns, what your promotional strategy should be, and how to manage the process across different AI engines. The downside risk is much smaller if you go early and get your learning in before it’s scaled.
Everyone talks a lot about cost saving and productivity with AI, but there are very few true cases that talk about top-line growth. Now, here’s a new channel that might generate new sources of growth and, particularly since there’s no front end to hassle with (as the front end is delivered by AI), margins might be higher
One example of a brand really exploring the potential of agentic commerce is Shopify, which partnered with OpenAI in September 2025. Shopify’s quarterly results, which will be released this week, will be some of the first clear signs of whether or not agentic commerce is taking off. Although it probably won’t touch the sides of their quarterly earnings in terms of revenue, many eyes are upon Shopify because they were one of the first, if not the first, to get some of their merchants online. Now, we’ve got one full quarter’s worth of information showing if that’s beginning to take off.
While it might not contribute very much to Shopify’s Q4 2025 results, this will affect the future value of their business. I think it is disproportionately important to the valuation of their business overall. So even if it’s not in the numbers, I expect this to be strongly there in the narrative.
What a platform like Shopify is doing is bringing AI capability to every mom-and-pop shop out there. And, if you’re a Fortune 500 retailer, you should be looking around and saying, “Well, these mid-market guys have now got the capability to go live immediately through their partnership with Shopify, with our in-house built platform, how ready are we to go?” That might put the blue chips at a bit of a disadvantage unless they’re developing at the same pace that Shopify has.
The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.
